Using the SpotifyR wrapper for the Spotify Web API, I collected various features from songs by the indie/hip-hop artist Guardin. The get_artist_audio_features() function of the SpotifyR package provides the user with various details about the requested artist’s music, such as album name, song duration, danceability, valence, energy, loudness, and many other features. Since these were the terms I focused my visualization on, the Spotify API defines the terms as such: danceability is a measure of the songs suitability for dancing, energy is the intensity/activity of the music, loudness is measures in decibels (dB), and valence is the percieved positiveness or hapiness conveyed by the track. Danceability, energy, and valence are all measured on a scale from 0.0 (lowest) to 1.0 (highest). After creating the heatmap, one of my key findings was that the energy of a song does not necessarilly equate to its danceability. This is shown in Guardin’s most recent album “Ataraxy Ii”, where the average danceability score was quite high, but the energy, loudness, and valence were all quite low. The albumns are arranged so that the most recent album is displayed at the bottom and his oldest album appears at the top of the heatmap. Overall, it seems that as the years have progressed, Guardin’s albums have become less “positive” and subdued in terms of overall sound. The wordcloud to the right shows Guardin’s most used words in his songs. By this wordcloud, it is easy to see that guardin uses alot more explicit language in his songs than I once thought, and talks alot about gloomier topics. A limitation I discovered when creating these visualizations is that my analysis does not really explore individual songs, as I though comparing albums would be more meaningful.